Prime Big Deal Days is your Q4 dress rehearsal. Treat it like one.
Profit Ninja Editorial
In a few days Amazon opens a 48-hour sale that Adobe expects to clear $9.9 billion in US online spend. That stopped being an Amazon story a while ago. Walmart, Target and Best Buy run their own deals the same week, discounts show up across retail, and roughly a quarter of the people buying are already shopping for gifts. For an affiliate it offers the one thing Black Friday never will: a real peak you can learn from while clicks are still affordable.
We wrote in August that summer is for testing and autumn is for scaling. October 6–7 is where those two modes meet. Here's how to use it.
What actually happens on October 6–7
The mechanics are simple. Two days, Prime members only, 35-plus categories, and new "Big Deals" dropping three times a day at midnight, 8 a.m. and 1 p.m. Pacific. Amazon describes the event in its own FAQ as "designed to kick off the holiday shopping season", and the numbers agree with the marketing for once.
Last year the two days pulled $9.1 billion of US online spend, up 7.3%, with discounts peaking around 18% off. The whole of October came in at $88.7 billion. For 2026 Adobe is forecasting $9.9 billion for the event and $95.8 billion for the month, and it keeps using the same phrase: an "industrywide e-commerce moment", with "discounts happening broadly across U.S. retailers".
Read that carefully. Those totals are all of US retail, not Amazon's till. If your offer sits anywhere near e-commerce, finance, apps or anything a person pays for with a card, your traffic is going to behave differently for 48 hours whether you planned for it or not. You may as well plan for it.
Why it's the rehearsal and not the show
Put the October numbers next to November's. Adobe's forecast for Cyber Week is $47.5 billion; Cyber Monday alone should do $15.1 billion in a single day, Black Friday $12.9 billion. Peak discounts run to 30% rather than 19%. Same shoppers, same muscle memory, roughly a fifth of the intensity.
That ratio is the whole point. The cost of learning rises with the intensity. Within's tracking, published by Jon Loomer, put average Meta CPMs at about $9 in October 2023 against $12 over the holidays, and in 2024 the jump was steeper: around $7 in early November to about $15 by early December. Triple Whale measured Meta CPMs at $22.26 across BFCM 2025. The exact figure moves every year. The direction doesn't. Every lesson you buy in October costs a fraction of the same lesson bought in late November.
One caveat worth carrying into your analysis: October is less mobile than the real peak. Adobe had smartphones at 51% of October 2025 spend, against 56% for the full holiday season and over 61% on Thanksgiving Day, with 63% forecast for this year. So an October read slightly under-weights your mobile numbers. Scale your mobile findings up, not down.
Four things to test while it's cheap
You don't need a dozen experiments. You need clean answers to four questions before CPMs move.
1. Angles, not just creatives
Deal-season traffic responds to framing: deal-led, value-led, urgency-led. Run your two or three leading angles against this crowd and see which one actually earns the click when every competing ad is also shouting about a discount. The creative testing framework and finding winning angles cover the method. The event supplies the audience.
2. The tracking chain, under load
Postbacks that fire at 200 conversions a day can misbehave at 2,000. Confirm your S2S postbacks before the spike, not during it, and run every tracking link through the Link Checker from the GEO you're buying, on a phone. A broken postback on October 7 costs you a day. On November 28 it costs you the quarter.
3. Landing page speed, on a phone
Half of this event is mobile and the real peak is more so. Load your page on a mid-range Android over a mobile connection, then strip whatever isn't earning its place. You'll find at least one script that is.
4. Your cash cycle
This is the test most people skip. October is the first realistic look at how fast money comes back to you after you spend it. Count the days from ad spend to payout in your account, because that number, not your ROI, will set how much of Cyber Week you can afford to buy. We pay affiliates on a 48-hour cycle for exactly this reason: a campaign that works shouldn't sit idle waiting for its own revenue. More on sizing that exposure in managing campaign risk.
Operator takeaway
Budget this event for learning, not profit. Spend enough to get statistically honest reads on angle, mobile conversion, postback match and cash cycle. If it also makes money, good. If it doesn't, you've just bought the cheapest version of a lesson you were going to pay for anyway.
The halo nobody warns you about
If you don't run Amazon offers you might assume this week is background noise. It isn't, and the data says the opposite in a way that catches people out.
Numerator surveyed over five thousand verified Prime Big Deal Days shoppers last year. 56% compared prices at other retailers before buying, most often Walmart (68%) and Target (43%). Over a third also shopped Walmart's Holiday Deals that week, and more than a quarter hit Target Circle Week. The deal appetite spills everywhere.
The trap is in how that spill lands on your numbers. Impact.com analysed over 1,300 non-Amazon retail brands across the July Prime Day: on brands' own sites, affiliate-tracked spend fell 45% during the four event days, but average daily spend in the run-up rose 14%, and the typical shopper journey stretched from 9.4 to 11 days. Demand didn't vanish. It pulled forward and compressed around the event.
So if you run non-Amazon offers, don't judge the 48 hours in isolation. Judge the window from the first of the month to a few days after. A dip on the event days is a redistribution, not a verdict. And remember who these buyers are: 84% of them told Numerator they expect to shop Amazon again for the holidays within three months, and 23% used the sale to buy gifts. The October customer is the November customer, a few weeks early.
How to read what you get
Five numbers are worth writing down. Each one in October predicts something specific about late November.
| What you measure | A good sign in October | What it predicts for Cyber Week |
|---|---|---|
| CTR by angle | One angle clearly ahead | Your lead creative. Build the seasonal variants from it, not from scratch. |
| CR, mobile vs desktop | Mobile within ~20% of desktop | Your landing page will survive the 60%-mobile days. |
| Postback match rate | Clicks and conversions reconcile within a few percent | Tracking will hold at five times the volume. |
| EPC against payout | EPC holds while CPMs rise | The offer has margin headroom when the auction heats up. |
| Days from spend to cash | You can state the number | Exactly how much Cyber Week you can finance. |
If you want a structured way to turn those reads into decisions, the campaign optimisation workflow is the loop we use.
The week after
By around October 12, with the Walmart and Target weeks also behind you, you should be able to write five lines on a page: your lead angle, your backup traffic source, the one thing that broke and how you fixed it, your real EPC at deal-season intent, and your cash cycle in days. That page is your November plan. It's also what you take to your affiliate manager, because the conversation about higher payouts, caps and private deals for Cyber Week goes very differently when you arrive with data than when you arrive with hope. When the numbers justify it, scaling campaigns covers how to add budget without breaking what's working.
FAQ
My offers aren't on Amazon. Does this event matter to me?
Yes, for two reasons. Walmart, Target and Best Buy run their own deal days the same week, and Numerator found 56% of Prime Big Deal Days shoppers compared prices at other retailers first, so deal-hungry traffic spills across retail, finance and apps. And the point of the rehearsal isn't Amazon: it's seeing how your creatives, tracking and cash cycle behave under a real spike while the lessons are still cheap.
Should I scale hard during Prime Big Deal Days?
Only what already works. Treat the two days as a rehearsal: modest budget, maximum learning. You want clean reads on angles, mobile conversion and postback match rate, not a bigger bill. Scaling is what November is for, and the October data is what makes that scaling safe.
What if my October results look worse than I expected?
Then the rehearsal did its job. A weak read on October 7 costs you a day and a modest budget; the same discovery on November 28 costs you the quarter. Fix the broken thing, re-test during the Walmart and Target weeks, and go into Cyber Week with a campaign you've already seen fail and recover.
Rehearse in October. Perform in November.
Prime Big Deal Days hands you a real peak at a fifth of the intensity and a fraction of the cost. The affiliates who treat it as a rehearsal walk into Cyber Week knowing their angle, their numbers and their cash. Everyone else finds out live, at $22 CPMs.
Sources
Amazon — Prime Big Deal Days 2026 FAQ (dates, mechanics, "kick off the holiday shopping season").
Adobe Digital Insights — October 2025 actuals ($9.1B event, $88.7B month, 51.4% mobile) and 2026 holiday forecast ($9.9B event, $95.8B October, $47.5B Cyber Week).
Adobe — 2025 holiday season recap (mobile share by day).
Numerator — Prime Big Deal Days 2025 shopper tracker (comparison shopping, competing sales, gift buying).
Impact.com — Prime Day halo effect analysis (July 2026) and holiday shopping trends 2026.
Within / Jon Loomer — holiday CPMs 2024 and 2023; Triple Whale — BFCM 2025 Meta benchmarks.
Pick the campaigns you'll rehearse with
Browse live offers across e-commerce, finance, nutra and apps, test them this week, and keep the capital moving with 48-hour payouts.
